Psycharakis, Stelios and Miller, S (2006) Estimation of errors in force platform data. Research Quarterly for Exercise & Sport, 77 (4). pp. 514-518. ISSN 0270-1367Full text not available from this repository. (Request a copy)
Force platforms (FPs) are regularly used in the biomechanical analysis of sport and exercise techniques, often in combination with image-based motion analysis (e.g., Begg & Kamruzzaman, 2005; Kuitunen, Komi, & Kyrolainen, 2002; Rahmani, Dalleau, Viale, Hautier, & Lacour, 2000; Rodano & Squadrone, 2002). Force time data, particularly when combined with joint positions and segmental inertia parameters, can be used to evaluate the effectiveness of a wide range of movement patterns in sport and exercise (Bartlett, Messenger, & Lindsay, 1997). According to Dainty and Norman (1987) and Bartlett et al. (1997), valid and reliable force measures depend on low threshold, hysteresis and cross-talk, high linearity, adequate sensitivity and the elimination of cable interference, electrical inductance, and temperature and humidity variations. Moreover, a platform must possess high stiffness and high natural frequency and be located such that extraneous vibrations are excluded.
Given that FPs are regularly used in sport and exercise research not only for data collection but also for evaluating other biomechanical equipment (e.g., Rahmani et al., 2000), the lack of attention paid to the likelihood of errors in their measurements is somewhat surprising. Although the scientific literature for kinematic data has established and consistently reported the estimation and propagation of errors, FP data are often taken as error-free. For example, Johnson and Buckley (2001) used a FP to measure ground reaction forces during sprinting, without calculating or reporting possible errors in FP data. Reporting FP data to an unjustifiably high precision and assuming they are acceptably accurate is potentially problematic. For example, when the force data are used for further calculations, such as estimating joint moments from external forces, the error in the force measures will propagate through the calculation--especially in the absence of postprocessing techniques--and directly affect the final result. Some investigators, however, have tried to assess the accuracy and reliability of their measurements. Bobbert and Schamhardt (1990) and Mita et al. (1993) found inaccuracies in estimates of the center of pressure position, especially toward the platform edges, and identified poor calibration and differences in the individual characteristics among the load cell amplifiers as possible causes.
Although FP calibration data are usually available from manufacturers, researchers should not assume the manufacturer-quoted values are retained following installation and over time. Experimental error is inevitable in a study that uses FP as a data collection tool and can arise from a variety of sources, which may influence the reliability of the findings and the study's validity. Despite this, neither calibration of FP nor estimation of potential errors in FP measurements is reported in most investigations. Furthermore, because the existing methods reported in studies for FP error calculation require sophisticated equipment and are time-consuming, their application in other FP studies would be rather complicated and, in many laboratories, not possible. Therefore, the purpose of the present study was to establish the magnitude of possible measurement errors from a FP typically used in sport and exercise biomechanics, by applying an easy-to-use, rime-efficient method.
|Uncontrolled Keywords:||biomechanics; force plates; kinetics; reliability;|
|University Divisions/Research Centres:||Faculty of Health, Life & Social Sciences > School of Life Sciences|
|Dewey Decimal Subjects:||500 Science > 530 Physics|
600 Technology > 610 Medicine & health > 612 Human physiology
|Library of Congress Subjects:||Q Science > QP Physiology|
Q Science > QC Physics
|Depositing User:||RAE Import|
|Date Deposited:||30 Jun 2008 12:01|
|Last Modified:||21 Mar 2013 11:58|
Actions (login required)